The Other Side of Automation

It seems to us that a magnificent new product made in Silicon Valley is announced every day in the press, on television, and via social media outlets around the world.  From an historical perspective, there’s no doubt that many of these innovations have improved the productivity of governments and the profitability of corporations, and they’ve enhanced the lives of everyday, garden-variety folks like you and us.  But we wonder if we aren’t approaching a point in time when advances in AI (artificial intelligence) and robotics will destroy jobs faster than the economy can replace them.

In the near future, self-driving cars, SUV’s, trucks, and buses will be cheap and reliable enough to be sold in volume.  Driving will be safer, and commuters will be more entertained and less stressed, but operators of commercial vehicles will be made obsolete because their vehicles won’t need them anymore.

According to the Bureau of Labor Statistics (BLS), there are 2.4 million commercial truck drivers in the United States who on average make about $50,000 per year.  In the near future, those jobs will be lost to automation.

The BLS estimates that 7.5 million retail jobs will be lost to automation.  More than half of retail workers are women who make about $23,000 per year.

A recent report by the McKinsey Global Institute forecasts that as many as 73,000,000 American jobs may be lost to automation by 2030.  If their forecast is in the ballpark, then the US economy will lose a number of jobs that’s approximately equal to 45% of the nation’s current workforce. (See Note 1.)  But the McKinsey report isn’t as foreboding as it would seem at first glance, because it also suggests that an overhaul of the economy could more than offset the 73,000,000 job losses.

Here’s the rub: the overhaul would have to rival or exceed the magnitude of our transformation from an agriculture-based economy to the Information Age, which took 165 years.  This time, though, we have twelve years, which means we have to move fourteen times faster than we did last time.  (See Note 2.)

We on The Other Side have a few questions:

1) Is the overhaul a training/retraining problem (as they suggest), a social problem, a political problem, a financing problem, or some combination?

2)  If politicians are even remotely involved, what are our chances?  Dodd-Frank wasn’t passed until after the Great Recession.  Our bridges, ports, railways, and highways are crumbling, but as of this writing Congress has yet to bring an infrastructure renovation bill to the floor of either house.  When was smoking in public places outlawed?  Was it before or after millions of Americans died from lung cancer?

3) Who’s going to pay for the retraining of the existing labor force: the public sector, the private sector, or the displaced workers themselves?  We expect it’ll be all three, but will government move quickly enough (see the above), will corporations be more disposed to retrain their employees than hire trained ones, and how much money will the unemployed have to invest in retraining themselves?

4) What’s the real-world scale of the problem?  How many bus drivers, bank tellers, and baristas will have to be taught to program computers, mine data, and repair robots, how long will it take, and what percentage of them will make the grade?

We could go on, but you get the point.

Unemployment during the Great Depression peaked at 25% in 1933.  In the absence of another cataclysm or two, we’re confident that the unemployment rate won’t come close to 25% in the next ten to twenty years.  But we’re just as confident that the pace of job destruction will far exceed the economy’s capacity to compensate for the losses––and no one will do anything about it until it’s too late.

At South by Southwest this year, Elon Musk said, “Mark my words: AI is far more dangerous than nukes, by far, so why do we have no regulatory oversight?  This is insane!”  (See Note 3.)


1) According to Wikipedia, 161,000,000 Americans held part or full time jobs in the US at the beginning of 2018.  Seventy-three million divided by 161 million is 45.3%.

2) One-hundred and sixty-five divided by twelve is 13.75.  We rounded up.

3)  The quote from Elon Musk was from ZDNet, a high-tech website.  We did not change a word, but we did edit the punctuation.

4) A roomful of statisticians and economists at the BLS could model the most probable outcomes, but they’d have to base their forecasts on so many assumptions that every politician, private-sector economist, and talk-show pundit could find at least five to dispute––for one news cycle.  The next day, they’d be arguing about something else.  We have two bits of advice for the BLS: a) forget the model, and b) tell your kids to get degrees in robotics.

5)  The Apex Child is a novel about this very problem.  Fair warning: it’s longer than this article.